chanchan mao (@ChanChan_Mao) header
chanchan mao (@ChanChan_Mao) profile picture

devrel, marketings, events @lancedb

sf
Joined June 2013
ok stepping outside for one reason only :) TONIGHT ‼️ @savant_vc is hosting @youliangtan , co-founder at @dreamlabs_ai, to talk about world action models, scaling robot learning, the next frontier of physical AI, and building at the frontier. RSVP: luma.com/joaw545m
i do NOT want to go outside today
1
108
grand prize winner 🥇 thank you @nebiusai, @nvidia, and the panel of judges for recognizing the work that we’re doing at @lancedb from the article: “More teams collect, clean, and test real-world robot data… LanceDB’s multimodal lakehouse speeds the data curation and feature engineering behind large-scale AI model training” what a great win after announcing Series B yesterday 🎉
Out of 363 physical AI startups, @nebiusai named LanceDB the Software, Tooling & Orchestration winner of its 2026 Physical AI Awards 🎉 Congrats to fellow winners Archetype AI, @Outsight_tech, @Luminary, and @ChefRobotics! na2.hubs.ly/H093DcK0
66
i do NOT want to go outside today
1
3
161
chanchan mao retweeted
Yesterday, we announced our $57.5M Series B. Today, meet LanceDB Realtime Engine. New AI data becomes searchable in milliseconds. In our tests, another client could read new rows a median of 52 ms after write acknowledgment. lancedb.com/blog/lancedb-rea…
2
5
66
5,020
chanchan mao retweeted
just shipped: an easy tool for posting your top 10 of 2026 pick your top albums, songs, movies, books or tv shows taylor.fyi/top
226
2,449
1,276
17,442
1,970,597
series B typa lunch 🤪
3
16
586
chanchan mao retweeted
Tell all your AI research friends that they should be building on @lancedb (if they're not already)! Very exciting things ahead for this company, and @BessemerVP is stoked to be leading the series B. To learn more, I highly recommend watching the recorded sessions from Reverie to hear how teams at @nvidia, @runwayml, @netflix, @coreauto, @AppliedInt, @ExaAILabs, and others are building on multimodal data at scale. Lots of other useful info on the docs page too - including tutorials and cool demos. Congrats to @changhiskhan, @eddyxu, and full LanceDB team.
The next generation of models needs a data stack built for AI. We just raised a $57.5M Series B to help researchers turn petabytes of video, images, and audio into better models, faster. Read the announcement: lancedb.com/blog/lancedb-ser…
1
2
13
593
It’s a great day to be a friend of Lancedb, it’s an excellent day to be an investor in Lancedb, it’s an incredible day to be a customer of Lancedb. People love to ask me my technical pov on what kinds of software and products will stick around through the wave of saaspocalypse, software factories, and personalized software. I have a variety of responses but I always end up saying that fundamental infrastructure like Lance which is already powering several of the largest training, serving, and improvement stacks in the world. The future is multimodal, and multimodal ai is build on LanceDB. It’s what dreams are made on.
The next generation of models needs a data stack built for AI. We just raised a $57.5M Series B to help researchers turn petabytes of video, images, and audio into better models, faster. Read the announcement: lancedb.com/blog/lancedb-ser…
2
3
35
1,498
what a huuuuge day for @lancedb 🎉 i joined the team a little over a year ago and i’m so amazed by how much we’ve grown - the team, the open source community, the adoption!! we even hosted our 1st summit this year! bringing together 250+ AI builders that all feel a similar pain: how to make the data flywheel spin faster for your next training run i’m incredibly proud of what we’ve accomplished so far, so honored to be apart of this team, and excited to see how much we’ll do next 🧡
The next generation of models needs a data stack built for AI. We just raised a $57.5M Series B to help researchers turn petabytes of video, images, and audio into better models, faster. Read the announcement: lancedb.com/blog/lancedb-ser…
6
1
20
632
chanchan mao retweeted
The next generation of models needs a data stack built for AI. We just raised a $57.5M Series B to help researchers turn petabytes of video, images, and audio into better models, faster. Read the announcement: lancedb.com/blog/lancedb-ser…
29
33
28
202
49,407
chanchan mao retweeted
finally some good ux for buying bananas online good job amazon 👍
65
39
82
2,624
649,209
working on building out the Reverie wing. first batch of session recordings dropped today btw! lancedb.com/reverie
1
56
???
Dual waterspouts on the San Francisco Bay seconds ago. I am in awe.
1
1
197
chanchan mao retweeted
I gave a keynote at Reverie where I covered the most important topic in building AI systems of today: love ❤️ and how to get that into space. youtu.be/tj9I3DrzNvQ For the next presentation If get invited, I will talk about purpose - if anyone wants to hear me yap.
Hi! this is Rohan Anil, I develop training algorithms among other things - my work trains various frontier models. At CoreAuto, I was fascinated by the problem of training on compute islands, eventually in space training. I hope you all enjoy this write up!🪐🛰️ coreauto.com/hitchhikers-gui…
11
8
3
156
9,872
random collection of latte art
3
229
chanchan mao retweeted
look, we get it — building good benchmarks is extremely hard...but that’s exactly why benchmark slop needs to be called out we notice this frequently from the company behind ParseBench, and ignored it the first few times, but don’t think its fair to customers, our competitors, or the broader community who look at benchmarks to make informed decisions selectively choosing datasets, gaming scoring methodology, or hastily throwing stuff together that’s riddled with errors just so you can appear #1 as a marketing tactic is disingenuous. While we tell users to eval on their own data, not everyone does and benchmark posts (even if flawed) shape opinions. parsebench heavily favors llamaindex, who created it, and like their other benchmarks has a ton of issues with it: - penalizes small formatting errors that users don’t actually care about - its overfit to their models, and introducing small amounts of variance and noise will dramatically change the results Extend has its own benchmark, RealDocBench, that we spent a lot of time and effort into creating: - we built it around agent Q&A, because that’s what users actually care about. nobody’s agent breaks because a heading was missing an underline or had a few extra spaces. instead, we ask whether an agent can answer questions correctly on parsed output from production-grade healthcare, real estate, finserv, etc documents - we did our best to make it fair. its why our competitors also do well on it (e.g. reducto’s r-1 performs well matching our performance parser, r1-mini matching light, etc) this is a massive market that all of us can compete in. long term, helping users make informed decisions is best for all of us it’s not worth sacrificing your credibility for short term marketing tactics
7
6
3
82
8,412